New prosthetic methods are giving people with motor impairments alternative communication and control channels. A logical culmination of these developments is a system that allows the brain to bypass completely its normal output pathways. Recent studies from this laboratory have shown that humans, including those with motor disabilities, can learn to change rapidly and accurately the amplitude of the 8-12 Hz mu rhythm in the electroencephalogram (EEG) recorded over sensorimotor cortex. Furthermore, they can use this control to move a cursor on a computer screen. Good single-channel control has been obtained, and initial data indicate that multichannel control is also possible. Thus, the mu rhythm, which recent work shows is detectable in nearly all adults, may support a multichannel brain-to-computer interface, and thereby provide a powerful new communication and control option for severely disabled individuals. This project's goal is a reliable multichannel brain-computer interface. The proposed approach is based on three well-supported hypotheses: that the scalp-recorded 8-12 Hz mu rhythm comprises a number of relatively independent components, that topographic analysis and frequency analysis techniques can distinguish between these components, and that humans can learn to control specific components and use them to operate a multichannel brain-computer interface. The first objective is to define by topography and frequency the separate 8-12 Hz components that are present when individuals are using the current interface. The second objective is to determine which components individuals are best able to control. The third objective is to incorporate these trainable components into a multichannel brain-computer interface that is rapid and reliable. These objectives will be achieved by combining online studies in which subjects learn to use the interface while extensive data are stored, and offline data analyses in which methods for improving the interface are defined. This project should produce an EEG-based brain-computer interface of significant value to individuals with disabilities. It should also lead to further work exploring the practical capabilities and theoretical implications of this new form of communication.